Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Status
The first post of this thread is a WikiPost and can be edited by anyone with the appropiate permissions. Your edits will be public.
Hi Chris,

I was able to track down an old version of the source code for ATY_init, from https://www.insanelymac.com/forum/t...amebuffer-xcode-project-to-start-from/page/3/

Build fails on 10A190 ppc but compiles file and kextload fines under 10.5.8 ppc but I guess the plist needs to be changed (doesn't recognise my PC Radeon HD cards under 10.5.8 - of note I can get them recognised but not working using https://forums.macrumors.com/thread...6-driver-radeonhd-framebuffer-driver.2344940/).

I also found the source code for ATY_HD (related), build also fails under 10A190 but compiles fine under 10.5.8; however it complains about IOACPIFamily when kextloading. Found it in the same thread.


Cheers,
This is a very interesting find indeed, thanks @pc297!

It may be worth trying against 10.6.8 and reporting on the development thread. If it’s found to work on 10A190 down the line can be written up here.
 
This is a very interesting find indeed, thanks @pc297!

It may be worth trying against 10.6.8 and reporting on the development thread. If it’s found to work on 10A190 down the line can be written up here.
I'll give the compiled 10.5.8 kexts a try under 10A190 and 10.6.8, however re ATY_HD my guess is that it needs older rather than newer frameworks, the error message under 10.5.8 was "22IOACPIFamily is not compatible with its superclass. 16IOACPIFamily superclass changed?"; so possibly frameworks/kexts from an earlier Leopard version - which could match that of 10A190, which is older than 10.5.8.

There are also these versions of Natit that were compiled for ppc if anyone wants to try them, and the old source code from 2006 (found at https://www.insanelymac.com/forum/topic/32004-natitkext-titan-imitation-with-source/ and on the above-mentioned thread)

Will write further here or on the 10.6.8 thread if I can get ATY_Init or Natit to work with either, with or without rom injection in their plists (the very last version has a few roms in there already)

Cheers,
 

Attachments

Just a quick question about using the nVidia 7800GS with SnowLeo PPC: I have one sitting in a box and in the past I have flashed it succesfully either with the "Tiger" firmware (temperature sensor OK with 10.4) or the "Leopard" firmware from MacElite site (temperature sensor probing removed). In the end I reflashed it with PC bios and abandoned in favor of a flashed FireGL X3-256 that at least under OS9 it support the screen resolutions chooser (but not, of course, 2D or 3D acceleration)
Now that I am having fun with SnowLeo PPC I would give to the nVidia card a second chance (from what I have read it seems to have a better support on 10.6 PPC than ATIs). My question is: if I flash the 7800GS with the standard firmware (temperature sensor functional) it will control the GPU fan nicely like in Tiger or it will full blast like in Leopard?
The PowerMac I am using is a FW800 dual 1.42 MHz
 
Hi,

quite interesting thread this one!
Actually I once had a patched 10.6 kernel booting to single user mode on a PowerBook G4 (or was it a G4 upgraded PowerBook G3?). Getting to the UI login screen wasn't possible because only the frameworks and libraries had ppc slices. It also didn't boot on an iMac G5 because some 64 bit ppc support had been stripped out of the kernel, maybe because Rosetta was 32 bit ppc only.

Leopard WebKit is another thing - getting the latest version working on 10.6 should definitely be possible. There have been attempts to get it working but the persons trying it lost interest when almost having finished the porting. And only the latest versions make any sense at all in my opinion.
For helping you I'll always need the complete build log attached here. It will take some weeks or even months to get it working - you'd really need to be very patient.
Or maybe someone could send me a usb flash drive that would boot into 10.6 on my late 2005 PowerBook G4? Or upload an image somewhere?

Best regards,
Tobias
I am currently working on your 604 LeopardPC version to compile against SL x86_64 , although I can compile it there are crashes which I try to fix as I go. the 537 SL version does not crash with Safari 10.0.5 (Webkit1 engine). What is the best way to approach the challenge?
 
I am currently working on your 604 LeopardPC version to compile against SL x86_64 , although I can compile it there are crashes which I try to fix as I go. the 537 SL version does not crash with Safari 10.0.5 (Webkit1 engine). What is the best way to approach the challenge?
Some crash reports might help to answer your question.
In case your goal is to run a WebKit2 version of Safari with a later WebKit, be warned because WebKit2 had changed that fundamentally after SnowLeopard, that I would strongly recommend staying with WebKit1 on SnowLeopard as well.
So my approach would be to first get running Safari 5.0.5 with the later WebKit and only after that try to port the later WebKit2 to the WebKit2 API which Safari 5.1.x did use - although I would definitely stay away from Safari 5.1.x .
 
I think Webkit2 on SL is a very tedious almost impossible job as most of the API's and system calls don't exist in 10.6 SDK, so yes the Webkit1 is the practical target. My question is if 604 is the limit or is there a way to compile the webkit for Catalina? Safari 5.1.x relies heavily on webkit2 that is why I don't even consider it (not that I did not try).
Many XPC calls that go nowhere in 10.6.
 

Attachments

Last edited:
Is this relevant to powerpc or is it x86 stuff out of the blue?
I have no way to test it on PowerPC, but it is compiled for x86-64 so pretty sure it is not related to powerpc. Even If I compile it for PowerPC I can't test it. But for PowerPC it is already compiled by @internetzel.
Safari 1.0.5:
1781397616243.png
 
Last edited:
  • Like
Reactions: ediflorianus
I think Webkit2 on SL is a very tedious almost impossible job as most of the API's and system calls don't exist in 10.6 SDK, so yes the Webkit1 is the practical target. My question is if 604 is the limit or is there a way to compile the webkit for Catalina? Safari 5.1.x relies heavily on webkit2 that is why I don't even consider it (not that I did not try).
Many XPC calls that go nowhere in 10.6.
Entirely possible to go further than 604, for x86_64 as well as ppc - I had a next version halfway done in fact (while half way means still a lot of work to be done yet).
This biggest problem is the missing TLS ciphers in the Security framework - which is why porting later WebKit versions seems far less attractive. Nowadays one could maybe use AI for getting full TLS 1.3 support into the Security framework. I guess you're using a proxy server to work around that problem?
 
  • Like
Reactions: barracuda156
Entirely possible to go further than 604, for x86_64 as well as ppc - I had a next version halfway done in fact (while half way means still a lot of work to be done yet).
This biggest problem is the missing TLS ciphers in the Security framework - which is why porting later WebKit versions seems far less attractive. Nowadays one could maybe use AI for getting full TLS 1.3 support into the Security framework. I guess you're using a proxy server to work around that problem?

What prevents us from using whatever modern libraries needed, as long as we talk about Safari? (Sure enough, we cannot make something like libSystem to depend on third-party libs, but why not applications.)
 
What prevents us from using whatever modern libraries needed, as long as we talk about Safari? (Sure enough, we cannot make something like libSystem to depend on third-party libs, but why not applications.)
It's "just" the integration work - and keep in mind that WebKit for MacOS uses OS API for doing TCP/IP transfers. Those libraries use other libraries and even kernel modules for doing the ciphering.
That's why I had patched and shipped an upgraded Security framework with Leopard WebKit - for at least supporting the minimum of TLS1.2 .
 
  • Like
Reactions: barracuda156
It's "just" the integration work - and keep in mind that WebKit for MacOS uses OS API for doing TCP/IP transfers. Those libraries use other libraries and even kernel modules for doing the ciphering.
That's why I had patched and shipped an upgraded Security framework with Leopard WebKit - for at least supporting the minimum of TLS1.2 .
How do I install this on Intel SL? It seems universal fat binary everything inside screams Snow Leopard
 
How do I install this on Intel SL? It seems universal fat binary everything inside screams Snow Leopard
For SnowLeopard the changes to the sources would have to be merged into the SnowLeopard version of the framework. The Leopard version is not compatible wirh SnowLeopard.
 
For SnowLeopard the changes to the sources would have to be merged into the SnowLeopard version of the framework. The Leopard version is not compatible wirh SnowLeopard.

It is something which you could do? (Not necessarily right now, of course.) It will be helpful for our powerpc project.
 
It is something which you could do? (Not necessarily right now, of course.) It will be helpful for our powerpc project.
I would need to know the base versions of quite some Apple open source projects (mostly all those with "security" or "Security" in their name).
In the end it would not be necessary to "sneak" in the patched framework, but it could directly be provided with the "SnowLeopard-PowerPC distribution", right?
But maybe that sneaking would still be necessary ... all those nasty hacks needed for getting this to work ... basically the SSL/TLS levels need to be remapped to later versions because the API doesn't know the later versions ... and at least in Leopard the updated framework used to break the system keychain, that's why it had be sneaked in only where actually needed (Safari, Mail, ...)
 
What do you mean "used to"? They're still there.
The sources for the Leopard Security framework have not always been completely available. Important parts of the code were from the OpenTransport library from the classic MacOS and released as open source only after Apple switched those parts to a newly implemented kernel module.
That way, suddenly the complete sources were available many years after the release of Leopard and a complete Security framework could be built from source where it wasn't possible before.
Maybe Apple could also do the opposite at times and take down source code from their servers.
 
  • Like
Reactions: barracuda156
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.